this... looks like a nightmare to maintain

@TheArcaneBrony what makes you say that? zod is great and looking through the diff, it looks much simpler. removes ajv entirely. I always hated things like these comments

I feel like those comments are perfectly fine? On the other hand, that specific case should be handled in code rather than schema validation... (ie. for use with 8 or 10 digit TOTP codes)

• need to run prettier
• need to fix the test ci
• should do a pass and simplify some of the schemas. the ChannelModifySchema for example, you've set all of the properties to optional rather than just using a .partial() on the parent object

The CI fails on main too so I feel it's kind of out of scope of this PR especially since it still fails after running the linter plus nix CI succeeds. I went through somewhat and did psrtials I'm at work though now so I can't do anything more.

I though ajv was unable to work with GW, if we could get something to work there, that'd be great

Zod should be able to work with gateway since it just parses and validates data, I was working on converting more stuff to Zod since this just did the API, I've got some partial progress but I feel that's another PR?

Hey, just gonna point this out now. I don't think this works at all. Zod does not have a way to do type cohesion everywhere which is required here.

Hey, just gonna point this out now. I don't think this works at all. Zod does not have a way to do type cohesion everywhere which is required here.

What do you mean we literally have z.infer giving us a ts type??

cohesion, not checking